What the Numbers Still Say—And What They Miss
Back in our 2023 blog on search engine rankings, we used DB-Engines as an impartial—if imperfect—way to measure the relative popularity of traditional search platforms like Elasticsearch, Solr, and Splunk. It offered a helpful snapshot of developer mindshare, particularly in open-source and enterprise contexts.
But in 2025, that lens feels increasingly outdated.
You can’t talk about search today without also considering:
- Databases acting like search engines – (e.g., MongoDB Atlas, Redis with vector search)
- Search engines evolving into data platforms – (e.g., Elasticsearch and OpenSearch now support analytics, observability, and vector indexing)
- The UI of search transforming – (blending keyword, vector, and GenAI-generated results into hybrid interfaces)
In short, “traditional search” can’t be evaluated in isolation anymore.
So, what does DB-Engines still offer in 2025?
- A useful signal of platform usage and long-term trends
- But a limited view of where search is actually heading
This year, we’ll explore:
- What the rankings still reveal
- What they miss entirely
- And which platforms are leading the shift toward hybrid, AI-powered search
DB-Engines Still Signals Developer Interest—But the Context Has Changed
DB-Engines still gives us a broad signal of which platforms developers are using and talking about. But in 2025, those rankings mean different things depending on how you slice them.
The rankings make more sense when viewed through two lenses:
- Search-first platforms – where indexing and retrieval are the product
- Embedded platforms – where search is just one feature in a broader data or application stack
Search-First Rankings
These are purpose-built engines designed specifically for indexing and retrieval—think website search, intranets, product catalogs, and internal dashboards. When search is your product, these are the platforms that matter most.
Elasticsearch still tops the list, followed by OpenSearch, Solr, and other traditional engines. Splunk is uniquely dominant in the log analytics / observability space, but not in broader semantic search use cases.
These rankings reflect developer mindshare—but mostly within the classic search engine space.
Search-Embedded Platforms
When you expand the scope to include data platforms where search is just one feature, a different set of leaders emerges.
In this view, MongoDB leaps to the top spot, followed by Redis and Elasticsearch—a lineup that reflects a new reality:
“Search is no longer siloed. It’s baked into your data layer, your AI stack, and your application logic.”
These rankings include multi-model and document databases where search is just one of many features—and that’s significant.
- MongoDB Atlas Search now rivals traditional engines, offering full-text search, scoring, faceting, and hybrid vector support.
- Redis is emerging as a fast, lightweight solution for semantic vector search at the edge.
These tools highlight a broader shift:
“Search is no longer a standalone layer—it’s embedded. Built into your data model, woven into your app infrastructure, and fueling GenAI pipelines.”
One notable footnote:
While Pinecone still ranks #68 overall, it has disappeared from DB-Engines’ vector and search engine subcategories in 2025. This likely reflects a shift in DB-Engines’ classification, not a drop in adoption. Pinecone remains widely used in GenAI workflows like retrieval-augmented generation (RAG).
So Which One Is “Best”?
That depends entirely on your use case:
If search is your primary interface, choose a dedicated engine.
Vespa.ai excels where large data sets and low latency are top requirements in traditional and AI Search (it powers Perplexity).
Algolia’s API-centric approach is very popular in e-commerce for product search and discovery.
If your application centers on data modeling, storage, or AI integration, then platforms like MongoDB or Redis may offer more long-term flexibility.
MongoDB stands out for blending flexibility with capability:
- Non-relational, yet supports relational-style joins
- Document-first, with robust full-text and vector search
- Cloud-native and built for GenAI and RAG use cases
“The real takeaway: don’t choose based on DB-Engines score—choose based on architecture fit.”
Platform Highlights from the 2025 Rankings
Search-First Platforms
Elasticsearch
Still the most recognizable name in search. Elastic remains a default for enterprise and observability use cases, now with hybrid vector support. It leads—but others are catching up. Elasticsearch is also widely deployed in observability and log analytics, so it’s leadership in semantic search use cases may be less pronounced.
OpenSearch
A growing open-source alternative backed by AWS. Good fit for users who want ELK-like functionality without licensing complications. Now supports vector search.
Apache Solr
A long-standing open-source leader, Solr remains in production across many search applications. But with newer AI-native workloads demanding hybrid and vector capabilities, it’s less commonly chosen for greenfield projects.
Algolia
Still a top choice for SaaS and ecommerce apps. Algolia is fast to deploy and easy to tune for user-facing search. Pureinsights has growing involvement with Algolia in support of rapid POC and frontend-focused implementations.
Coveo
Coveo delivers AI-powered search and recommendations, with strong support for semantic search, relevance tuning, and personalization. Widely used in ecommerce and support, it combines vector search, NLP, and user behavior signals to improve content discovery. Pureinsights has partnered with Coveo on select implementations.
Vespa.ai
A new entrant to the DB-Engines list, Vespa is open-source and designed for real-time, AI-native search and recommendation. Pureinsights is proud to partner with Vespa on high-scale, GenAI-ready deployments.
Splunk
A leader in log and machine data analysis, Splunk isn’t a pure semantic search engine—but it incorporates semantic elements like its Common Information Model, AI-powered SPL assistant, and deep learning for text similarity to support more context-aware queries.
Pinecone
Though dropped from DB-Engines’ vector/search lists in 2025 (now ranked #68 overall), Pinecone remains a go-to for GenAI pipelines like RAG. It’s specialized for vector-only retrieval and designed for scale and simplicity.
Search-Embedded Platforms
MongoDB
MongoDB’s Atlas Search has matured into a true hybrid search solution—combining document storage, full-text indexing, and vector search. In 2024, the company acquired vector database startup VoyageAI, reinforcing its commitment to GenAI use cases. This move strengthens Atlas Search’s position as a go-to platform for teams building AI-powered content discovery and retrieval applications.
Redis
Redis is now more than a cache. Its vector search modules support low-latency, edge-friendly semantic search in GenAI applications.
Couchbase
Quietly capable multi-model database with full-text and vector search support. Ideal for teams looking for flexible deployment models.
Databricks (Honorable Mention)
While not ranked as a search engine, Databricks is gaining relevance for search-related workloads in enterprise AI projects. With growing support for embeddings and retrieval inside Lakehouse architectures, we’re starting to see real-world traction.
Where We See Search Heading in 2025
“Your architecture—not a popularity contest—should drive your search platform decision.”
Final Thoughts: Don’t Build on Rankings. Build on Fit.
The DB-Engines list still has value—but it increasingly feels like a rearview mirror in an era driven by hybrid, AI-powered architectures
It tells us who’s busy. But it doesn’t tell us who’s breaking new ground.
In 2025, the best search experiences are:
- Semantic and contextual
- Vector-enabled and hybrid-aware
- Embedded in your stack—not bolted on
- And often orchestrated as part of AI pipelines, not just keyword queries
Whether you’re building search for a website, an intranet, or a GenAI assistant, the platform you choose should reflect your architecture, your users, and your future—not just a score.
We hoped you got some value out of our Search Engine Rankings 2025. Curious what hybrid or GenAI search could look like in your stack? Let’s talk —or request a demo to see it in action on your own content.
Additional Resources
- Discovery 2.2 + Vespa.ai – Unlocking Hybrid Search at Scale – Pureinsights
- Smarter Search with Discovery 2.1: Inside Our Voyage AI Integration – Pureinsights
- Guest Blog: Search Platforms vs Vector Databases for AI Search – Pureinsights
- Elasticsearch vs OpenSearch in 2025: What the Fork? – Pureinsights
- DB-Engines Search Engine Rankings – June 2025
- DB-Engines Rankings with Secondary Models Enabled
- Overall DBMS Rankings – June 2025